arXiv:2610.08984
Valley transitions during spin shuttling in Si/SiGe heterostructures
Jonas R F Lima, Guido Burkard
Abstract
Valley excitations are considered one of the biggest challenges for a coherent spin shuttling in Si/SiGe heterostructures. In this work we theoretically investigate the valley dynamics during spin shuttling in a Si/SiGe heterostructure by considering a two-level system given by the two low-lying valley states of the mobile quantum dot. We assume that the electron is initially in the valley ground state and analyze the evolution of this state during shuttling. We find that, depending on the shuttling velocity, the system can be in three distinct regimes. In one of these regimes, the quantum information is preserved even in the presence of the valley excitations, allowing for coherent spin shuttling independently of the shuttling distance. This regime is obtained for a shuttling velocity of tens of m/s, a desired regime where dephasing mechanisms, such as spin-valley hotspots and spatially varying spin splitting, can be suppressed. This result is in accordance with recent experimental work that obtained a high spin shuttling fidelity after a fast and long shuttling. We also propose strategies based on the Partial Landau-Zener model to enhance the valley shuttling fidelity, which was found to be much more efficient than previously proposed strategies based on the standard Landau-Zener model.
AI Summary
The paper asks how disorder-driven valley transitions affect spin shuttling in Si/SiGe, and whether control should target valley-splitting minima or changes in the complex intervalley-coupling phase. It calculates position-dependent intervalley coupling from an effective-mass model with randomly distributed Ge atoms, then numerically evolves the two-valley Schrödinger equation. Simulations use several dot sizes and interface widths, typically over 300 nm, with disorder averages over 1,000 realizations. Three velocity regimes are reported: slow shuttling with distance-dependent valley leakage; intermediate velocities with sequences of nearly complete transitions between instantaneous valley eigenstates; and faster shuttling with sustained valley superpositions. Representative intermediate-regime velocities are approximately $0.1$–$10\,\mathrm{m/s}$, with parameter-dependent boundaries. Both faster regimes can yield mean ground-valley occupation near $1/2$, despite different individual dynamics. The paper argues that complete transitions can preserve spin information, producing a deterministic spin rotation, provided valley relaxation is avoided and the electron finishes in the ground valley. Control proposals include selectively slowing one transition, adapting velocity to the valley-phase gradient, and lateral path optimization. Phase-adapted schedules achieve mean valley fidelities $0.987$–$0.994$ at mean velocities $6.9$–$42\,\mathrm{m/s}$, versus $0.665$–$0.968$ at $4.6$–$6.4\,\mathrm{mm/s}$ for the tested splitting-based schedules. The evolution model omits spin, so spin-coherence conclusions are inferred rather than directly simulated. The paper identifies experimental implementation of phase-informed velocity and path control as challenging; its proposed coherence-preserving regime remains conditional on avoiding relaxation and arranging the final valley state.
Central claim may not hold: The spin-free valley simulations do not establish the central assertion that full valley transitions avoid spin–valley entanglement and preserve spin quantum information. The abstract omits the body's qualifications that valley relaxation must be avoided and the electron must finish in the valley ground state, an endpoint obtained in only roughly half the uncontrolled realizations. The detailed results, abstract, and conclusion give inconsistent velocity ranges for the favorable full-valley-transition regime and the unfavorable combined-valley regime. Short individual excited-valley residence times do not bound cumulative relaxation over long transport, so the presented finite-distance results do not justify distance-independent preservation. The stated Landau–Zener transition formula has an exponent four times too large for the Hamiltonian and parameters displayed.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.